aboutsummaryrefslogtreecommitdiffhomepage
path: root/tensorflow/python/grappler
Commit message (Collapse)AuthorAge
* Consolidate device parameter arguments into a shared DeviceInfo structGravatar A. Unique TensorFlower2018-10-08
| | | | PiperOrigin-RevId: 216280197
* Updating the V2 variables API.Gravatar Alexandre Passos2018-09-27
| | | | PiperOrigin-RevId: 214824023
* Open source graph analyzer.Gravatar Yao Zhang2018-08-27
| | | | PiperOrigin-RevId: 210439649
* fix C++ header guards.Gravatar A. Unique TensorFlower2018-08-21
| | | | PiperOrigin-RevId: 209679086
* Assert closeness of output values instead of equality in layout optimizer testsGravatar Smit Hinsu2018-07-19
| | | | | | | | using convolutions Convolution output may differ across convolution algorithms and is not guaranteed to match exactly. PiperOrigin-RevId: 205276671
* Merge commit for internal changesGravatar Mingxing Tan2018-06-20
|\
| * Reduce Grappler overhead by skipping optimizers when the graph is tiny.Gravatar A. Unique TensorFlower2018-06-18
| | | | | | | | PiperOrigin-RevId: 201095811
| * Automated g4 rollback of changelist 201011811Gravatar Akshay Modi2018-06-18
| | | | | | | | PiperOrigin-RevId: 201033171
| * Merge changes from github.Gravatar Akshay Modi2018-06-18
| | | | | | | | PiperOrigin-RevId: 201011811
* | Potential fix to layout_optimizer_test.pyGravatar Michael Case2018-05-31
|/
* Remove _USE_C_API staging in tests now that the C API is enabled by default.Gravatar Skye Wanderman-Milne2018-05-16
| | | | | | This is in preparation for removing the _USE_C_API toggle altogether. PiperOrigin-RevId: 196920481
* Optimize idempotent ops, e.g., Snapshot(Snapshot(x)) => Snapshot(x)Gravatar A. Unique TensorFlower2018-05-04
| | | | PiperOrigin-RevId: 195308675
* Simplify file reading and support SavedModel.Gravatar Yao Zhang2018-05-03
| | | | PiperOrigin-RevId: 195291836
* Mark all nodes processed by AddOpsRewrite/MinBCast stages with a tag.Gravatar A. Unique TensorFlower2018-05-02
| | | | PiperOrigin-RevId: 195167597
* Use the default rewriter config instead of a custom oneGravatar Benoit Steiner2018-04-30
| | | | PiperOrigin-RevId: 194824761
* Run 2 passes of rewrites by defaultGravatar Benoit Steiner2018-04-26
| | | | PiperOrigin-RevId: 194443770
* Merge changes from github.Gravatar Yifei Feng2018-04-23
| | | | PiperOrigin-RevId: 194031845
* A single measurement is enough when using simulationGravatar Benoit Steiner2018-04-06
| | | | PiperOrigin-RevId: 191934781
* Enable constant propagation across Enter nodes, but only if is_constant is true.Gravatar A. Unique TensorFlower2018-04-04
| | | | | | Don't propagate constants with control dependencies through Merge nodes. PiperOrigin-RevId: 191663396
* Fix a shape inference bug.Gravatar Benoit Steiner2018-04-03
| | | | PiperOrigin-RevId: 191528009
* Apply "Raise exception in SWIG on bad TF_Status" to base.iGravatar Akshay Modi2018-04-03
| | | | | | Minor fixes to make this work. PiperOrigin-RevId: 191457070
* Fixed the interaction between virtual cluster and measuring cost estimator.Gravatar Benoit Steiner2018-03-27
| | | | PiperOrigin-RevId: 190712404
* Added an option to run shape analysis assuming the shapes of the feed nodes areGravatar Benoit Steiner2018-03-21
| | | | | | valid. PiperOrigin-RevId: 189923541
* Don't inline functions in the grappler item builder since this part of the ↵Gravatar Benoit Steiner2018-03-16
| | | | | | | | code doesn't support custom ops. Instead we will rely on the function optimizer. PiperOrigin-RevId: 189400462
* Remove underscore prefix from ref_identity op.Gravatar Anna R2018-03-14
| | | | PiperOrigin-RevId: 189096108
* Automated g4 rollback of changelist 187582263Gravatar A. Unique TensorFlower2018-03-02
| | | | PiperOrigin-RevId: 187657654
* Automated g4 rollback of changelist 187563544Gravatar Gunhan Gulsoy2018-03-01
| | | | PiperOrigin-RevId: 187582263
* Grappler: Change memory optimizer recomputation name prefix into a regexp. ↵Gravatar A. Unique TensorFlower2018-03-01
| | | | | | | | | This allows us to match any node names, especially those under different scopes. This still performs a prefix regexp match, so it is basically backwards compatible. PiperOrigin-RevId: 187563544
* Sampling group embeddings for each child.Gravatar Rui Zhao2018-03-01
| | | | PiperOrigin-RevId: 187532388
* Removing underscore prefixes from hidden generated Python functions.Gravatar Anna R2018-02-28
| | | | PiperOrigin-RevId: 187386941
* Fix print format error.Gravatar Rui Zhao2018-02-26
| | | | PiperOrigin-RevId: 187033623
* Preserve user placement as much as possible when optimizing the graphGravatar Benoit Steiner2018-02-23
| | | | PiperOrigin-RevId: 186822511
* Measure the performance of the original placement to ensure that we preserve itGravatar Benoit Steiner2018-02-22
| | | | | | in case the placer isn't given enough time to find a better solution. PiperOrigin-RevId: 186655094
* Started to open source the RL placer.Gravatar Benoit Steiner2018-02-21
| | | | PiperOrigin-RevId: 186563773
* Ensured that the model pruner outputs the nodes of the optimized graph in aGravatar Benoit Steiner2018-02-21
| | | | | | deterministic order PiperOrigin-RevId: 186520272
* Temporarily disabled part of a test that fails on MacOSGravatar Benoit Steiner2018-02-21
| | | | PiperOrigin-RevId: 186494795
* Make sure the nodes that are refered to by a collection are preserved during anGravatar Benoit Steiner2018-02-20
| | | | | | optimization PiperOrigin-RevId: 186394467
* Shorten grappler per-node report by default with previous behavior behind ↵Gravatar A. Unique TensorFlower2018-02-20
| | | | | | --verbose. PiperOrigin-RevId: 186369380
* Support multiple fetch nodes and add a flag for memory report.Gravatar Yao Zhang2018-02-20
| | | | PiperOrigin-RevId: 186329308
* Made cost_analyzer_tool accept fetch nodes when running with metagraph ↵Gravatar A. Unique TensorFlower2018-02-16
| | | | | | option. Also made it read metagraph in either binary or text format. PiperOrigin-RevId: 186010810
* Automated g4 rollback of changelist 185072479Gravatar A. Unique TensorFlower2018-02-15
| | | | PiperOrigin-RevId: 185900165
* Fix a bug to update reduction axes for all supported cases. Turn off layoutGravatar Yao Zhang2018-02-14
| | | | | | optimizer for all reference runs, as it is on by default now. PiperOrigin-RevId: 185782777
* Support reduction with true keep_dims and squeeze along NHW dimensions.Gravatar Yao Zhang2018-02-12
| | | | PiperOrigin-RevId: 185411786
* Do not convert layout for Select if condition input is of unknown shape.Gravatar Yao Zhang2018-02-10
| | | | PiperOrigin-RevId: 185242138
* Only convert format if input is of layout-agnostic type.Gravatar Yao Zhang2018-02-08
| | | | PiperOrigin-RevId: 185103227
* Add memory usage report to cost analyzer tool; run all default optimizations.Gravatar Yao Zhang2018-02-08
| | | | PiperOrigin-RevId: 185058999
* Set the number of warmup steps for building the cost model.Gravatar Yao Zhang2018-02-07
| | | | PiperOrigin-RevId: 184878186
* Support parsing from text and fused op in contrib.Gravatar Yao Zhang2018-02-05
| | | | PiperOrigin-RevId: 184558131
* Enable aggressive identity node pruning in dependency optimizer.Gravatar A. Unique TensorFlower2018-02-05
| | | | PiperOrigin-RevId: 184539756
* Disable graph optimizations (CSE) in test so that constant nodes are not ↵Gravatar Yao Zhang2018-02-02
| | | | | | deduped. PiperOrigin-RevId: 184289685